Abstract

Enhancing Performance Tracing and Debugging in Remote Deployments: Leveraging In-Memory Asynchronous Logging and trace log analysis with Google Cloud Platform (GCP)

Cloud computing and containerization have emerged as dominant paradigms in modern software development, enabling standardization, portability, efficiency, and scalability. However, remote deployments pose unique challenges, particularly in ensuring traceability and facilitating effective debugging when the away machine or virtual machine (VM) is not directly accessible. The constraints of setting up asynchronous log emission mechanisms further compound these difficulties.

This presentation will discuss the challenges of remote deployments and how in-memory asynchronous logging and trace analysis can be used to address these challenges. The presentation will cover the following topics:

  1. The foundations and principles of in-memory asynchronous logging;
  2. The benefits of in-memory asynchronous logging; analysis of logs - traceability, queryability, and performance measurement through trace aggregation ;
  3. A case study on the use of this technique and trace modeling, extraction, analysis and automation with Google Cloud Platform (GCP)

The presentation will conclude with a discussion of the outcomes and performance improvements observed in the case study. Problems such as tracing journeys, hotspotting and concurrency, file system IO issues and memory leaks and performance trends which were solved using this technique will be discussed.

Attendees will gain a comprehensive understanding of this technique and its relevance to remote deployments. They will learn about the practical implementation of this technique and how it can be leveraged to achieve traceability and performance measurement in real-world scenarios. Additionally, insights into the integration with GCP will be shared, enabling attendees to explore the synergy between these technologies.

Biography

Alankrit Kharbanda and AJ Ortega are tech leads at Google.